Picture from a Lochinvar WHN-55 installation

smihaila

Hi everyone,

My father and I have recently finished a DIY heating project, using Lochinvar WHN-55, Biasi steel panel radiators, Grundfoss Alpha 15-15 pump and copper piping.

I have posted detailed pictures if you are interested.

Thank you to all the members of the Heating Help community, who provided us with very wise insights and they were of tremedous help in deciding how we have approached the project.

smihaila

Details

-House computed heat loss: 35 kBTU/h

-Location: Ontario, Canada

-Boiler: Lochinvar WHN-55

-Connection scheme: directly coupled hydraulic circuit, only one pump.

-Boiler/System pump: Grundfoss Alpha 15-55

-Rads: Steel panel rads from Biasi. Oversized a lot, in such way that we can work with a 130F / 55 degrees Celsius as design temp. It's enough for the rads to get a bit warm, and you are enjoying a bath of warmth.

-Pump debit: 3 gpm.

-ODR and system/well sensor installed.

-Thermostat: EcoBee SmartThermostat

-Piping: 1'' iron near piping, 3/4'' "L" copper for main trunk and 1/2 "L" copper for the rest. Piping scheme: parallel.

-No automatic filling valve. Just the manual one.

I have also some electronic thermostatic valves from Danfoss (Eco), but haven't installed them. I'm uncertain now that they are needed.

smihaila

More pictures

Set #2

smihaila

One more set

The third set.

Thank you.

hot_rod

some creative piping

in that installation. Are you running the Alpha on a fixed speed setting? I like that Ecobee stat. I put on in my shop last year, great features.



Wise choice on the Caleffi DiscalDirt

smihaila

Space

was limited, so hence the creative, tri-dimensional piping :-)

The only thing we weren't sure about the setup, was the placement of the expansion tank - according to Extrol docs, it's supposed not to be put at a "deadend pipe", because presumably it would cause excessive corrosion. But I'm sure what's considered deadend pipe and which not.



The Alpha pump, yes, I run it at fixed speed, level I. It draws only 7 watts. If I run it at medium or dP-const, it would move too fast, at about 7-8 gpm. 3gpm in our case is satisfactory and quiet.



The ecobee is nice, indeed. I also have the remote sensor add-on board and it's too bad that a same thermistor/temp sensor like the one from ODR, or from the system temp well, can't be "shared" between the Lochinvar and the input module from ecobee. It's because everything is powered from the same source/the Lochinvar's 24 VAC transformer. I guess I'll have to buy separate sensor to leverage that module.



The Discal is also doing a good job of de-aerating. I was looking for an insulation shell from Caleffi but nothing is in stock or any US store willing to ship it to Canada :-). Ecomfort.com has it listed but they can't get it in fact.



About the pump, I was searching a lot for one which is not only ECM but also accepting a 0-10VDC control signal for var speed. But couldn't find anything besides the Magna or the Stratos non-eco model. Which are over-sized for our needs. It would've made it a nice setup, because Lochinvar could've controlled it directly. Eh, it's still good, and 7 watt makes me very happy.



We aren't pros, but we are stubborn and keep our traditions :-) Have you noticed the hemp? My father swears by it :-)



Thank you.

hot_rod

??

Is the Alpha pumping into the Knight? If so then it looks like the expansion is connected to the supply coming out of the boiler? Then you are pumping at the PONPC.



Is that pressure/ temperature gauge on the outlet, supply of the boiler? If so does the pressure increase or decrease when the pump starts?



Is the Alpha on speed 1 moving enough flow thru the heat emitters. I've found those GPM numbers on the Alphas to be a bit off sometimes, when I install a flow meter in the piping.



Watch the delta T across the boiler S&R when it is running, I think Lochinvar shows a 30 or 35F max. Make sure the single pump is moving enough flow when all zones are calling. You may need to bump to speed two.

smihaila

Correct

Your observations are correct. The Alpha is pumping into the boiler, and the tridicator and expansion tank are connected on the boiler's outlet/supply line.



I know, it sounds like a violation of the "pumping away" principle. I had followed the Lochinvar's diagrams to the letter, and I was always puzzled as to why you would connect it that way. But someone on this forum shed a bit of light on this matter. Seems that it's related to the fire tube design of this boiler. It likes to have the water pumped into it, from bottom to up. It also has something to do with some "presostat" which would get triggered if you connect the pump the other way. So, the conclusion was that we should not blindly try to follow the "pumping away" principle, when the boiler has other more stringent/particular requirements.



Believe me, I was divided between whether to do it that way or not. But it works fine.



The Delta-T reported by Lochinvar's SMART panel shows 12.5 Celsius/20 F.

I'll try to observe if the pressure in the gauge increases or decreases, I haven't paid attention too much. It has to be very insignificant.



The Alpha's reported GPM, yes, it can be with a large tolerance. Very rarely, I can see it at 4gpm instead of 3, but after a while it stabilizes at 3. The boiler is happy with it. I intentionally altered its settings momentarily, to make it fire for a longer period i.e. 15 minutes. And I have raised the ecobee to a very high target temp. And I could see how the setpoint temp was reached in the boiler (even though the thermostat was still calling for heat), and how the boiler started to modulate down, from 100% down to 90-85%. So I think it works within its normal/expected parameters. The flue's temp and the water temp is always in the expected range.



If I increase the Alpha's speed, it will go to say, 7 to 10 gpm, and at that point I will see the delta-T in the boiler getting low, very significantly. Like 5 or 10 Celsius degrees instead of 12.5C/20F. We have about 100-110 liters of water in the whole circuit, that's about 25gallons. The head/pressure loss seems to be low, as we made sure to use 3/4 L copper piping for the main trunk and then 1/2 L copper for the terminal lines. 1/2 L copper is significantly more debit than a 3/8'' pex or 1/2'' pex, right? I was afraid that lots of elbows and tees would create more loss, but I could blow with my mouths through the system and wasn't hard. Also, this Alpha doesn't seem to be so fragile/weak as I was afraid initially. Grundfoss folks really have a product to be proud about!



Thank you.

smihaila

Zone

[forgot to mention that it's a very simple setup - no multi-zone, no zone control valves, no zone controller, etc] The house is relatively small (1400sqf without basement, 1800 with basement included) and insulation is quite good. The thermal loss got even lower after installing triple-glazed windows from Germany, with an Uw factor of 1.1-1.2.

hot_rod

I think you are fine

Yes to the Alpha pumping into the boiler.

The DiscalDirt is in the best location.



If anything, I'd like to see the expansion tank connected at the inlet side of the Alpha. On speed 1 it will not develop much delta P so not a big issue. Happy, efficient heating!



Im still wrapping up my Lochinvar Cadet installation. I got it in heating mode today, 50 gallon buffer/ separator, ODR dialed in. The first cycle was boiler on for 25 minutes.

Eastman

pumping into whn-55

What exactly is the reason for pumping into the vertical fire tube?  More turbulence?



It's kind of funny to see a system actually setup up to run on one tiny pump.

smihaila

Don't have

an exact explanation (someone with more experience on fire tubes from the forum could chime in, or I'll try to find the exact post which was relevant to me). But I guess it has to deal with the considerably higher volume of water which this boiler has?



Yeah, it's very surprising to see this tiny pump working on a 2-floor house + basement. Applying intuition here is failing us, what counts the most is the momentum not the height of the column of water that needs to be moved/pushed upwards.



One thing that doesn't seem to correlate is this: At 20F delta and 3gpm, UHF says that the boiler would produce 30KBTU on the "conveyer belt". Assuming that my house heat loss calculation was correct (33-35k) and that the WHN-055 is supposed to be able to deliver 55kBTU max (at 100% fire modulation), then how come in my case the boiler reaches 100% modulation within 2 minutes of fire? I guess this should be seen as a transient phase, when any boiler would fire at max in order to convey the heat that was displaced/missing very fast. And the truth of the UHF formula gets confirmed only if the setpoint is matched in the boiler, BEFORE the thermostat ceases to call for heat? Only in that case I was able to see the boiler starting to modulate down, to 90%. Or it could be that the heat loss computations were too optimistic? Mind you, the original forced air furnace which was installed by the builder of the house (in 1990), is 40 KBTU. So, somehow the hvac-calc calculations were not that far off.

smihaila

Expansion tank

Nice job you did with the Cadet. Very clean and structured installation, congrats. 25 minutes for the 1st cycle wasn't bad at all.



About my tank - interesting, do you think that by mounting it on the inlet/suction side of the pump, the pump will work less hard?



The Installation manual can be found here:

http://lochinvar.com/_linefiles/WH-I-O-Rev%20M.pdf



If you check Page 35, Figure 6-1A, we can see boiler pump on the inlet, while pressure gauge and tank is on outlet side. But let's say it's not clear because both 6-1A and 6-1B figures involve a P/S coupling.



Page 38, Figure 6-3: shows the system pump indeed with the tank on suction side. This would be for a P/S coupling, and what you recommended is 100% in sync with that.



But...if we look on Page 45, Figure 6-10 - that's the diagram we've used, for a direct coupling, only boiler pump, no system pump: You can see the boiler pump not changing its place (it's the same as in P/S circuit) - so it's pumping into the inlet side. Then, we see the outlet going into the gauge first and then into the expansion tank. And the radiators/heating circuit coming right after the expansion tank.

I thought initially that someone a mistake has been done by Lochinvar. But on the next page, Page 46, Figure 6-11 we can see a variation of the direct connection scheme - multi-zone with zone valves and single pump.



The page 47, Figure 6-12 was the most interesting one: Shows a third type of direct connection, but this time multi-zone implemented with multi-circulators. In that case only, you will see the zone pumps connecting on outlet and AFTER the expansion tank. Similar to all the other P/S connection schemes. What's interesting in that 6-12 figure is that the zone pumps are no longer pumping directly into the supply line of the boiler, but they push into the heating circuit first.



For a direct circuit, and no zones, it looks like Lochinvar knows some secret to the recipe. Because besides pumping into the inlet side, also the tank is mounted on the opposite side, on the supply.



Thank you.

hot_rod

a glitch in the I/o manual

on page 39 Single Temperature. The boiler pump and System pump are in series, doubling the head, probably causing excessive velocity in the piping. I think that loop needs to be closed at the bottom and a P/S is developed with a system pump and a boiler pump.



Page 45 is called Full Flow, this looks correct, piping with a single pump, which is doable with the fire tube type HX. But I'm not sure the best location for the expansion tank connection with a system pump and DHW pump.



Lochinvar is aware of the error and correcting it.



Pumping into the HX, even a wide open style shows the delta P as a pressure increase, if you pump away from the PONCP. This helps with air elimination, and assures any pressure switches on the boiler are held closed. If it were a high head circ you would want to be sure the added head doesn't cause a relief valve to seep.



Years ago all cast iron package boilers were shipped with circs on the return. This wasn't necessarily a bad thing had they indicated the correct expansion tank connection at the inlet of the circ, not on the discharge, hanging off the air scoop and the circ pumping at the PONCP.



B&G of course knew this, as indicated in this 1966 schematic from B&G. Siggy added the color to it for Modern Hydronic Heating 3rd edition. It's much easier to follow with the two colors for static fill pressure and pressure added via a circ. But it clearly shows the importance of "Pumping Away"

bill_105

Nice install and nice Pics.

But after searching all the Pics. I still can't find Dear old Dads hemp.:)?

Eastman

I'm not sure I follow you...

The modulation behavior sounds generally correct.  The boiler's goal is to hit the the ODR set point basically as fast as it can, and then adjust itself to match the heat loss from the system.  Unfortunately the thermostat is ending the call for heat just as the boiler is beginning to modulate.  Try lowering the set point and perhaps take advantage of the ramp delay feature. 



Now, are you saying the delta T never varies form 20 degrees as the boiler changes modulation?

smihaila

Page 39 - P/S with zone valves

You are right, Figure 6-4 seems to have a mistake. They forgot to draw the P/S coupling (via closed spaced tees or LLH) right above the boiler circulator. Page 40/Fig 6.5 and all the other P/S schemes will show it ok.



Now, back to the Full Flow Figure 6.10 on Page 45. You are doubting about the location of the Expansion tank, when a  system + DHW pump is used? But the schematics is about not having any system pump at all. Also, in my case I don't have any DHW tank nor DHW pump.

You also mention that Lochinvar "is aware of the error and correcting it." You mean that they know that the tank should've been placed also on the inlet side, right after the radiators circuit returns and getting into the suction side of the boiler pump? Or I've read you wrong and you are referring back to that P/S based scheme (6-4)?



"Pumping into the HX, even a wide open style shows the delta P as a pressure increase, if you pump away from the PONCP." => Ok, so you are saying that pumping into the HX is not a mistake. And only the tank is placed incorrectly.

"This helps with air elimination, and assures any pressure switches on the boiler are held closed." => Are you saying that thestuff with the pressure switches is ensured not due to having the pump pumping into the HX, but rather through the fact that the tank will be placed in sync with the PONCP principle?

" If it were a high head circ you would want to be sure the added head doesn't cause a relief valve to seep." => Are you suggesting that this would happen in my case, with the tank placed wrong? Unintentionally higher pressure would build up possibly creating an opportunity for the 30 PSI relieve valve to trigger?

Or the fact that the added head is an inherent consequence of a "pumping away" type of connection (more dynamic pressure building up on the push side of the pump)?



"Years ago all cast iron package boilers were shipped with circs on the return. This wasn't necessarily a bad thing had they indicated the correct expansion tank connection at the inlet of the circ, not on the discharge, hanging off the air scoop and the circ pumping at the PONCP." => So, you are suggesting that Lochinvar in fact didn't do their due dilligence and have some exact reason for suggesting the placement of the pump to pump into the HX, and they just "inherited" some thing of the past and didn't want to bother with modern techniques like PONCP and do their calculations on their side more exact?



Please explain in layman's terms:

1. If pumping into the HX is a stupid assumption from Lochinvar.

2. If Page 45 has also a mistake as far as expansion tank is concerned, and if I am in real danger with keeping that tank the way it is. I have absolutely no issues with the air. There is a small air vent also placed at the highest point/2nd floor - on one of those towel rads.



There is lots of info on this forum, and hard to disseminate info properly (aka read between the lines). You try to do your best and follow all of your manufacturer's recommendations and schematics (otherwise the warantee becomes void), and you also watch closely all the past revision versions of the same I&O manual document, to make sure the info didn't get changed suddenly in some schematics (thus suggesting an obvious mistake), TO REALIZE at the end that you did it wrong? How can a manufacturer be so iresponsible and stupid? Or it's done intentionally to steer the DYI people away? :-)



If we wish to get into details, even when we think about the (static) filling circuit, I was at great debate with my father - he was saying that in fact the filling should be done from the return to supply side of a rad, which is the opposite direction of the normal heating flow. Because the air would escape much easier when you fill from bottom to up. But then we noticed that you guys here in the Northern hemisphere prefer to do the charging in the same direction as the normal flow, maybe because of that automatic filling valve thing, so that the filling could be done with the system fully working.



I'm a bit puzzled - we tried to do the things by the letter of the book, to find out that the manufacturer is either stupid or doing that on purpose. I must admit that my father told me about many situations whereas a manufacturer does it on purpose, in order to bring their men at work. It happened to him once with a wood/pellet-fired boiler back in Europe. But he realized himself all the mistakes and in fact people from the manufacturer came to him to see the install and took notes and learned from him how to do it right !



Thank you.

smihaila

Great idea

"Try lowering the set point and perhaps take advantage of the ramp delay feature."

=>Ok, I'll try that. I had the ramp delay feature enabled, because I wanted to have the last/more permanent stage of those 5 stages to last more than 10 minutes (which seems to be the default value that I observed when ramp delay is disabled).

Good idea to lower the set point/water temp, just to see the boiler modulating down but somehow letting the heat in the emitters accummulating more and more, to the point that the thermostat will get finally satisfied.



"Now, are you saying the delta T never varies form 20 degrees as the boiler changes modulation? " => I have clearly noticed the delta T changing (increasing) when the heating cycle starts. Now, I must admit I didn't pay attention to delta T when the boiler goes down from 100% to say, 90%. The logic says that it should still keep it constant/20F, because that's what is it aiming for, right? It strives for a constant water temp in its mini-tank, and if the temp rises above the target, or notices that the return gets near to supply, it modulates down or stops firing completely.

smihaila

Pressure check

I checked the pressure gauge carefully. In standby / pump off it's about 19 PSI. And when it's on, I see a noticeable bump, to 19.5 PSI. Then it stays constant. When turning off again, I don't see it dropping, at least immediately. Perhaps after a while will get back to 19.

Is this pressure increase expected / a good thing? Or, with the tank vessel placed before the pump, the correct thing to see is a decrease?

smihaila

Hemp

Can be seen on the iron piping, some excess on several threads. I've cut/hedged it a bit in the meantime, that's why maybe the recent photos no longer shows it more clearly :-)

Eastman

enabled?

"I had the ramp delay feature enabled, because I wanted

to have the last/more permanent stage of those 5 stages to last more

than 10 minutes (which seems to be the default value that I observed

when ramp delay is disabled)." 



Why is it going to 100% if the ramp feature is on?

smihaila

Ramp delay

goes up to 100% because I've left the modulation value per each stage to default. I customized only the time/period:

20% step 1 for 30 second

45% step 2 for 30 second

60% step 3 for 30 seconds

80% step 4 for 30 seconds

100% for the rest/permanent stage, for max 20 minutes.

When the setpoint was reached (by moving the thermostat target temp to a very high value on purpose), I was able to see in stage 5 how it started to go down from 100 to 90%. Maybe I can tweak those settings more.

I guess it's a trade-off: how long you wish the gas to be on (longer if you modulate down) vs how fast you want the heating / less time to be on.

Truth be told, I never undestood how the valve modulates. Is there a linear correspondence between the modulation value and the cubic meters of gas consumed? Or in fact the valve is always open and the modulation is done via controlling the air flux (which creates a sort of negative pressure/vacuum, which triggers the gas to be absorbed from the utility company at a higher rate/debit)?

Thanks.

Eastman

I don't remember all the features

of the ramp delay and the max output cap, but I believe you will want a configuration that prevents the boiler from firing above the max heat loss of 30 to 35k.  The boiler doesn't realize that this full high end output is unnecessary.  I would drag out the ramp up as long as possible, with the final output capped at approximately 35k.  100% high fire brings the system up to temperature quickly,  but full output is less efficient than longer burns at partial modulation and also maintains a steadier temperature in the home.



The gas is under pressure,  the more the valve opens the more gas will come out.  The fan operates in step with the valve, providing the required combustion air.

Henry

Creative

1) Why a Rinnai AND a Lochinvar?

2) A pump must 5 dimensions before and three after of straight pipe!

3) A street elbow in NOT allowed on gas piping in Canada. I am a voting member of B149 and B149.2

4) Has TSSA checked off on the install? Who was the licensed contractor that did the install?

Henry

More

Where you have the street elbow and bushing (illegal) is that a cast T above? Then that too is illegal. I assume the copper line is fuelling the hot air system or another appliance. Also code states that when you go down you need a drip tee/sediment pocket and not up into the appliance. Is that Teflon tape on the gas joints? If so, it is illegal also. We ban Teflon tape because a thread of it can prevent a gas regulator or valve from closing properly!

hot_rod

Mistake

In the I&O is the first drawing as you mentioned



Pumping into the boiler is fine. Learn more about PONPC and pumps in Dans wonderful book titled "Pumping Away". The schematics show it but it is good to read Dans explanation. With multiple pumps you need to think where the best spot for the tank connection is. Dan cleverly explains all that

smihaila

Some answers

1)Why a Rinnai AND a Lochinvar?

The Rinnai was installed 2 years ago, and at that time I wasn't planning to have hydronic heating. It's our first house.

2) A pump must 5 dimensions before and three after of straight pipe!

The space was limited, so not too much room for that :-(

3) A street elbow in NOT allowed on gas piping in Canada. I am a voting member of B149 and B149.2

Ok, we'll take galvanized street elbow down. Thanks for pointing that out.

smihaila

Other answers

"Where you have the street elbow and bushing (illegal) is that a cast T above? "

The Tee shown above the street elbow is an iron tee. Came from the original Rinnai install. What it should be instead? Is there a different type of iron tee to be used? Is the bushing also illegal? A reducing nipple should go then instead.



"I assume the copper line is fuelling the hot air system or another appliance."

Yes, it goes into the forced air furnace. The valve is turned off now for that trunk.



" Also code states that when you go down you need a drip tee/sediment pocket and not up into the appliance."



Yes, we have the drip tee mounted on all of the three gas apliacense: one at Rinnai, one at Lochinvar, and the other at the forced air furnace. There was another drip cap mounted right below the main gas line, but we took it out and placed there that pesky street corner :-)



"Is that Teflon tape on the gas joints? If so, it is illegal also. We ban Teflon tape because a thread of it can prevent a gas regulator or valve from closing properly! "



I'm not sure if it's teflon. Any white stuff that you see is from the original Rinnai install. Done by Conestogo Mechanical. Are you implying that Conestogo didn't follow the proper regulations? Any of our new intervention is using hemp, exclusively. No teflon. The father doesn't like teflon because: (1) you pay royalties to some damn Dutch company who invented it and (2) it's not long-lasting and it "affects the pipe's thread" my father so tells me. So it's hard to do further interventions on it with teflon.

SWEI

DIelectric unions

are not necessary in a closed loop heating system.



Pumping into firetube boilers should not be a problem (unless the pump is massively oversized.)

smihaila

@Eastman, thanks

for the explanation on how the modulation on the gas valve works. I know that seems logical, but somewhere on other trhreads (or other sites) someone was suggesting that there isn't any modulating gas valve per se and that the combustion is regulated ONLY via the air flow. Now your explanation make a lot more sense.

Besides a better combustion efficiency, another advantage which lower modulation brings is the pesky fan. I kind of dislike hearing it at full throttle (6,000 rpm) and I'm afraid this is quite a fragile piece in the Lochinvar scheme of things.

Ok, my heat loss calculation is a 30,533 BTU/h more exactly. I shall provision for 35k, so 35k/55k would give a max 63-64% modulation. Isn't that a tad small? :-)

I did more measurements at full 100% modulation, to see how the flow rate and delta Ts corelate with the max BTU which the boiler is advertising (55kBTU):

-Pump set at lowest speed, 3GPM, 7watts =>dT = 12-13 celsius, that is 22-23F. This makes it 33,750 BTU total.

-Pump set at medium speed, 7GPM, 24watts => dT = 7-8 celsius, that is 13.5F. This makes it 47,250 BTU total.

-Pump set to max speed or const dP (no matter what dP setting/min-med-high), 10GPM, about 45-50watt => dT = 5-6 celsius, that is 10F. Which is 50,000 BTU total.

I find the 3GPM vs 7GPM results interesting. At 3GPM one would expect the dT to go higher, in order to achieve the 55K. But doesn't seem to happen. It is as if "magically" the whole system adapted itself to the heat load of the house?? I don't believe in magic - there must be an explanation. If the 3GPM flow rate is insufficient to leverage the heat producing capabilities of the source, we should see a higher delta T or the boiler should modulate down (while dT is kept constant). The third option - BTU difference going somewhere else - is impossible. So I guess the explanation lays in that you have to discount the time until the ODR/setpoint is reached. And only then you would see the modulation going down. So, is it ok to keep the circ @ 3GPM and lower the ODR + set the ramp delay for stage 6 (the last) @ 66% modulation?

7GPM would be too high, above the max flow rate spec'ed by Lochinvar (5gpm is the max @ 20F rise, 4gpm @ 25F rise and 3gpm @ 35F rise). But I find it quite interesting - it's like the boiler listen to "UHF corelation", in the sense that, when the flow goes to 7GPM, the delta T will decrease accordingly, so the boiler is "protecting itself" and dT doesn't stay fixed, at 20F but goes down to 13.5F. Then, the question, why those tables with max flows are being published? :-)

Btw, with my current setup (3gpm / 22.5F delta and 100% modulation) I can see the boiler condensing very visibly. I guess the outdoor temp is not that low and this favorizes the condensation. It's for the 1st time when I can see it very clearly - was never able to see it with Rinnai (which runs @ 200 kBTU).

Lots of things to know, a fascinating domain. I'm an amateur, no doubt about it, but it's curious to see how many people in the trades are in fact "feeling" what's going on and have a mathematical justification for everything which is going on with the hydronics :-)

smihaila

Dielectric unions

So, it's not a danger of creating a Copper-Iron miniatural electric battery without them?



To be honest, the copper/sweat connections took us by surprise when we bought the Lochinvar. We were expecting to see iron piping (Lochinvar reps, if you see this, it would be great if you could mention this detail in the manual). Had we known that in advance, we would've mounted only copper for the near piping. It was too late to change the plan :-) Those dielectric unions (Taiwan made) don't look too solid at all, they may not last long.



Thank you SWEI (also for confirming the pumping into HX).

Eastman

the gpms

Are you reading the flow rate from the alpha?  I believe those can be quite inaccurate, people have commented on this in the past.  The lower flows probably have the biggest error.  You could very well actually have about 4, 8, and 10.5 gpms; giving you roughly 50k for each measurement.  Did the fan sound the same at each 100% modulation test?  That would be an indicator that it was firing at the same output each time.



35k is not small for something that's going to be running almost every minute of the day on the coldest day.  You'll have to play around with the ODR.  I don't have a methodology for picking just the right set points.  In a sense, the more perfect the setpoint, the less responsive the system is to user changes at the thermostat, wind, losses of other sources of heat, etc...  so as a necessity it must be somewhat higher than required.  You also don't want it to short cycle when the heat loss the home is below the boiler's min mod of 11k.

SWEI

Galvanic corrosion

decreases as pH increases.  Oxygen-free water (unless it started with an abnormally low pH) does not require the dielectric break that fresh water does.



FWIW, a threaded Schedule 80 PVC nipple or a section of PEX provides a far more effective break than a dielectric union does.

smihaila

Galvanic corrosion as factor of pH

Interesting, yes it makes sense, how I couldn't realize that? Lower the pH is, less acidic the water will be. And a more alkaline solution would decrease the corrosion effect.

Looks like I went to the national chemistry Olympiad during my High school years for nothing :-)



But mind you, it seems there are 5 groups of metals, responding differently to different pH values. I found this:

http://www.scribd.com/doc/84964563/7/Galvanic-Electromotive-Force-Series



Please check Page 40: "5. The last group is large: beryllium (Be), aluminum (Al), copper (Cu), zinc (Zn),cadmium (Cd), tin (Sn), lead (Pb), cobalt (Co), nickel (Ni), zirconium (Zr),gallium (Ga) and indium (In). They corrode both in acidic and alkaline solutions [...] These metals are resistant only to neutral solutions"



About the unions, yeah, it's one of those cases where too much of caution is a bad thing :-).



The folks back in Europe (where the copper is en vogue now and Pex-Al is no longer used much) never put them. It was a new thing for my father to see these DIelectric unions :-)



Thank you.

SWEI

Alkaline water

almost always comes from calcium and magnesium, both of which are buffers.  At least that's what I recall of my limited P-chem during engineering school.

smihaila

OMG, gpm readings

Eastman, you were so right about the possibly large variance for the Alpha's readings.

I did more tests and set the Ecobee in test mode, to allow the boiler to fire for as long as I like.



I've let it run for about 30 minutes, after which the Alpha suddenly started to shift its reading from 3 to 4 gpm and very shortly from 4 to 5gpm. And it was registering a stable solid 5GPM after that! What's the reason for such variation? Does it take longer for the Alpha's internal tachometer to get good average values? Or the pump just need time to get its bearings "fluid" or the whole system needs time to get into some harmonic state?



About the fan noise, yes, for the same 100% modulation factor, its noise seemed to be the same (at 6000 rpm) between a test to another.



When I've set the water set point to 45 Celsius/113F, I was able to see the boiler ramping down after sufficiently long time, to as low as 48%, and still condensing.



The only thing which I still don't understand is this: Why is the Delta-T linear with the modulation rate. I could notice this when both ramping up and down, so the result is very consistent and reproducible: At 100%, Delta-T is approx. 20F. At 50%, it becomes 10F.

Is it as if the boiler realizes that it's hard to control the gas modulation more precisely, and it tries to drag as long as it can on such decision, and instead it tries to be smart and reduce instead the Delta-T in order to achieve a lower BTU output? But if that's the case, then how come the fan also goes down? From 6000 rpm to about 3000pm.



Perhaps the explanation is much simpler: At a lower modulation, the rate at which the heat absorbed/dissipated by the rads becomes higher than the rate at which the boiler can produce and deliver the heat. And hence, the Temp differential between supply and return would naturally shrink?



About the ramp delay feature, btw I told you that I initially enabled it not to limit the upper modulation, but for a much trivial reason. Without the ramp enabled, some earlier tests that I've done lead me to the conclusion that the boiler would never fire for longer than 10 minutes, due to some allegedly default params in the Lochinvar's software. But no, on this occasion I disabled the ramping and with the Ecobee in test mode (call for heat ad-infinitum) I was able to see the boiler firing even longer than 20 minutes. I guess it was the ecobee playing some tricks during the initial tests. Ecobee has a sort of "smart recovery" feature, which I presume leverages some P-D/anticipative features and while calling for heat, will suddenly decide to stop the heat and then wait for a while / about 1 minute or longer. And then it would re-check the room temperature and if it's the case it will re-fire. So, it's like it wants to provision for some time for the radiant/convection heat to propagate in the whole space. I believe this feature (for which the time cannot be customized) interferes with the boiler's anti-cycling feature (which is set to block any consecutive calls for heat that are distanced at less than 1 minute).



Finally, I have this dilemma and obsession: the gas consumption and total number of hours when the boiler is firing. Is firing at 100% for 5 minutes equal to firing at 50% for 10 minutes, in terms of gas consumption? What about the "wear and tear"/stress posed to the boiler's burners and tank? Isn't it bad to have it firing longer, even if its a low fire?

I know that the combustion efficiency gets better at lower modulations, but what the sweet spot between gas consumption, efficiency and boiler longevity/stress minimizing would be??



My settings right now are:

-Max setpoint (ODR-independent): 70 Celsius/158F

-ODR water temp at +10 Celsius/50F: 45 Celsius/113F

-ODR water temp at -10 Celsius/14F: 65 Celsius/150F



I've set that 158F max value as a "worst case scenario" - in case when the outdoor temp goes crazy low (low than 14F) and 150F water temp wouldn't be enough to satisfy the Ecobee's call for heat.



If I recall properly, I designed for system design temp = 130F/55 Celsius.

Obviuously I need to fine-tune those params more.



My ramps settings are:

Ramp 1 of 6: 25% for 30 seconds.

Ramp 2 of 6: 40% for 30 seconds.

Ramp 3 of 6: 55% for 30 seconds.

Ramp 4 of 6: 75% for 1 minute.

Ramp 5 of 6: 100% for 4 minutes (to give the water temp a sort of "boost").

Ramp 6 of 6 (the permanent regime): 75% for max 20 minutes.



Does anyone have other magical values to recommend, which would be more optimal for my setup (35K heat loss, family liking more heat such as 75F and a Canadian SW Ontario region)?



Many thanks.

smihaila

Explanation for delta-t linear with modulation

I'm answering my own question. I believe I realize what's the explanation is.

With the BTUs getting lower with decreasing modulation, if the flow rate is constant, then logically, something has to give, and that can be only the Delta-T.



I was blind and I can see now. The very same low 10F delta-T at 3 GPM and 50% modulation factor is the same as in the case with 7-10GPM and 100% modulation factor.

So, when the output power lowers, if we wish to keep the same delta-t of 20F constant (which is preferable), we should lower the speed of the "conveyer belt". Otherwise the low firing power cannot keep up with the speed of moving the heat units to the rads.



Now I realize that a const delta-T pump (and with attachable temp sensors), OR 0-10V variable voltage control pump (which Lochinvar would be able to directly modulate) would have been the ideal. Now, since a small ECM pump with 0-10V control is a rarity, the only viable option should've been either a non-ECM/triac base varspeed pump or, for energy efficiency but rather more indirect control, an ECM const dT pump. Can't recall, is bumble bee const dT or const dP like Alpha?



The conclusion would be that, as far as maintaining a wider delta-T, in my case it's better to keep the modulation as high as possible, since we can't vary the pump's speed dynamically depending on the production rate or on the heat consumption/absorption. That's kind of funny, because you wouldn't go for high modulation just for that purpose only.



What a sucker I was, maybe I shouldn't have focused on low power consumption and go instead with a non-ECM varspeed? Ideally, what I wanted initially was to buy the bumblebee, but my heating dealer was not having it here, in Canada.

smihaila

pH

(sorry, I made a mistake, I meant the opposite - higher the ph is more alkaline would be).



Speaking of Calcium or Mg, yes indeed, those should make it more alkaline. Especially in an area with extremely hard water such as mine. But guess what, we've filled it with softened water, so no chance to get much alkalinity from that :-) At least, we tried a prudent approach (lot of debate exists about what water is the best for filling a closed loop system, especially one non-steam based, even on this forum - what's the lesser of these 2 evils: calcium precipitating on the copper tubing but no rust, vs. no precipitate but possible internal corrosion due to in-balanced ions/salt).



Our approach was to not fill with the softest of the softest water possible (i.e. right after a new regen cycle took place), but we have waited for the resin bed in the softener to come to about 50% usage. The softener is pretty new, but didn't wish to take any chance with any potential salt lingering on the resin bed.

Eastman

you're hung up on a non-existent problem

The efficiency of the system is not directly related to the deltaT.  A system with a constant 20 degree delta is not in and of itself desirable.  As you realized, a fixed flow rate combined with variable burner output necessitates variable delta T.  This isn't a bad thing, it's a consequence of a good thing. 



"The conclusion would be that, as far as maintaining a wider delta-T, in

my case it's better to keep the modulation as high as possible, since we

can't vary the pump's speed dynamically depending on the production

rate or on the heat consumption/absorption. That's kind of funny,

because you wouldn't go for high modulation just for that purpose only."



Maintaining a high modulation output is not the answer.  Efficiency, in the context of this discussion, is driven mainly by the temperature of the return water and the burner modulation rate.  Increasing the modulation rate would increase the delta T, but not by lowering the return temp.  Increasing the allowable modulation rate increases the delta T during the transient phase by raising the supply (output to rads) temperature.  It would not lower the return temperature.   A delta T pump is capable of lowering the return temp by reducing flow --basicaly the water is sitting in the rads longer allowing more time to cool.



Disregarding deltas and temperatures, high modulation rates are inherently inefficient.  High output gets a building up to to temperature quickly, but like driving a car, speeding from point A to B saves time at the expense of *more* fuel.  The high temperature gases in the combustion chamber move faster with increasing modulation, leaving less time for heat transfer to occur, the end result is a higher proportion of thermal energy escaping unused out the flue.



You can indirectly lower the return temps via the ODR setpoint settings, but not to the same extant as direct delta T control.  However, I am not aware of any delta T pumps or boiler controllable pumps that operate below 7 watts.  One would have to add hydraulic resistance to the circuit to bring the pump within it's range of authority at a higher wattage.  As an alternative, consider adding a hydraulic circuit setter to your system to enforce a precise fixed lower flow.  Or, call it mission accomplished.  The system is as efficient as it can be without adding further cost and complexity. I'm sure one can not find a more efficient boiler set-up for miles.

Eastman

inquiring minds want to know

what is the status of your system???

smihaila

well,

Yesterday I've finished painting all the black piping, in a pleasant tremclad gray colour.



The system is working fine, and I re-enabled the ramp delay feature, this time with the following settings: stage 1 = 60% for 10 minutes, 2 = 65% for 10 minutes, 3 = 70% for 10 minutes, 4 = 75% for 10 minutes, 5 = 80% for 10 minutes and as last resort, last stage 6 = 100% for 20 minutes. The ecobee thermostat seems to be satisfied at stages 1 or 2. You were right, it's better to drive the firing more slowly, to favor higher efficiency. I can notice that floating ball in the condensation drain -that its level went down a bit more - may be a clue that the boiler is condensing at a faster rate, not sure.



Also, did more tests with the pump. It seems that the flow rate self-increase (from 3 to 5 GPM) is not time corelated, but water temp corelated. Only when the boiler is about to reach the target set point, you will see it at 5GPM. I wonder what's the reason for that. It's as if more pressure builds in the system at higher temps ? But why isn't the tridicator reflecting that? Or, another possible explanation is that the boiler itself is somehow adding some head loss on purpose, as a safety measure? And in the end it decides to remove that additional friction?? It puzzles me. It would've been nice to install some sort of professional flow meter (which you could engage only at will, via a valve).

Also, maybe I'll have to rent a combustion analyzer from somewhere - not being a pro, I don't own anything of that sort. Based on your past experience with the boilers, notably Lochinvar, do you think that the manufacturer's settings are not precise, or the gas pressure at any location may be different than their assumption in the factory?



Overall, you have a satisfied customer here :-)



Thank you for explaining so well how the delta-T works with the boiler, and how a const delta-T operating pump would do it instead.

I've learned a lot from this forum, thank you to all of you for helping!